schemes

simulated data

Histogram of simulated abundance data (per year and region). The used scheme is identical - no trend.

Abundance per year and region. The used scheme is identical - no trend.

Model output

Data was simulated with identical - no trend, 20 sites per region (A and B) and 10 years in total. Each year, only 20 sites were included in the analysis (10 per region). For the second year (Y = 2), the ratio of A:B ranged from 0:20 to 20:0, the target ratio would be 1:1. Each ratio was repeated 20 times, each repetition randomly selected new sites. For each round of varying ratios from 20:0 til 0:20, the randomly selected data of other years remained the same (resulting in 20 different datasets which were each used for different ratios). The model formula was N ~ s(Y, by = R) + R, with N = abundance, Y = year, R = region, family = Poisson. Weighted models additionally added a weight to N. The respective weight for A (or B) in year 2 can be calculated as 0.5/(No. of A (or B) sites in year 2/20), the number of B sites is 20 - No. of A sites in year 2.

annual parameters

Only year 1-5 are included with year 2 being the most important comparison as this is the only year where the regions got different weights.

Standard error

Standard deviation

Model fit (median abundance)